Dr Su will present Zen 2 and Navi at Computex 2019 (May 27th)

[Stefan Payne]

8 hours ago, Neftex said:

what are you talking about? theres ton of leaks on zen

You know that you can spread false information and make up some shit to see where its leaking and who leaked it??

 

That's common practice these days and in the last couple of years, AMD has fixed all their leaks.

So the "Leaks" are often not leaks but Speculation. The Information that was released early was rather useless. The early samples were also clocked at around 3GHz or so. 

 

In the end, we didn't know much about the performance until it was officially released.

 

Same with Zen2. What do we know about Zen2?? 
The things we know about Zen is what AMD presented at CES earlier this year and also talked about. Not much more besides that.

 

OK; there was another Leak from a University or something like that, that showed that Rome has even higher AVX Performance than the Competition. 

And that's about it.

[Neftex]

@Stefan Payne dude youre asking for confirmed leaks before the product is released, ofc that doesnt exist. it wouldnt be a leak then.

 

there is a ton of leaks on zen2 but until the product is here ofc we wont know if theyre true or if theyre made up...

 

lastly, it doesnt matter what company the leaker is from. if he has correct information it doesnt matter if he hacked it off some pc or knows a source inside

[Master Disaster]

8 hours ago, JoostinOnline said:

I never said (or suggested) that. I don't know either way. Neither do you, and you should stop pretending otherwise.

Well you're the only person here naive enough to not believe it happens. Ftr Linus has stated that he knows it happens publicly in the past, I guess he's making up bullshit as well.

 

Fun fact: just because something cannot be proven as true doesn't mean it isn't true.

[ravenshrike]

41 minutes ago, Master Disaster said:

Well you're the only person here naive enough to not believe it happens.

The question is not does industrial espionage happen. It is whether Intel's sources were/are good enough to have the precise details that Moose is claiming. Given their reactions it's pretty clear that at best they only knew a month ahead of time for Zen and Threadripper as to the details of any announcements, and probably closer to a week beforehand. It's also pretty clear that the leaks on the current architectures are all coming from manufacturing partners or people much further down the chain at AMD and that AMD is at least trying to be tightlipped about specifics even internally. The reason you didn't see a similar pattern for the initial Zen launch was because of how tight the launch window was. Remember all the initial difficulties in supply that resulted from the tight turnaround time motherboard manufacturers had.

[mr moose]

11 hours ago, JoostinOnline said:

So your proof is that I can't prove it doesn't happen? XD

 

I'm not the one making claims. You're spouting Internet rumors and passing them off as fact. I'm not saying they're definitely false, but claiming they're definitely true is silly.

I have my head in the sand for calling out your appeal to ignorance fallacy?

Given that many people in the industry have said it happens,  there are actual legal cases of it happening.

 

https://www.themarysue.com/intel-information-theft/

https://www.therichest.com/rich-list/10-of-the-most-infamous-cases-of-industrial-espionage/

 

Not to mention I said I witnessed it first hand.  But I am hardly going to throw ex employers under the bus.

 

7 hours ago, Stefan Payne said:

No, it makes your statement false.

 

Or are you accusing Intel of Espionage?

 

How long did it take for the 6 Cores??
That doesn't look like they were prepared or even knew about it.

Especially with the fab problems, wich is in part due to the more cores and thus larger dies and also designed for maximum utilization of the Fabs with the smaller 4 Core Dies.

...or they didn't know more than we did and were only projecting and expecting something that wasn't as good as AMD delivered...

 

And thus they released even larger cores to the Consumer, wich they didn't want to and plan, wich also ruined their fab projections and caused utilization beyond 100%.

Ähm, no.

Intel didn't know until it was too late.

Intel made their fab for their projected sales of small 14nm/4 Core Dies and some server stuff.

 

Now Intel can't ship, even had to respin a Chipset in the older 22nm process to free up space in the 14nm Fab(s), so that processors can be done instead.

WIth what we know, its absolutely certain that they didn't know more than we did or not much more.

 

Same with nVidia as well.

AMD also didn't know that, if they did, they wouldn't have fired the one that made the VEGA 20 on Desktop...

 

And there are NO LEAKS on AMD's side that are noteworthy, wich then in turn means that they closed their leaks...

 

 

The only source of Information for the other company might only be Partners like MSI, ASUS and co, who might (or might not) get the Chips for valuation purposes.

 

So let me get this straight, If I am accusing Intel of espionage you agree but if I am accusing AMD of it then I am making false statements?  Typical.

 

Majority of what you said is just you not liking reality again.

 

[mr moose]

9 minutes ago, ravenshrike said:

 It is whether Intel's sources were/are good enough to have the precise details that Moose is claiming.

And what precise details where they that I supposedly claimed?

 

 

[ravenshrike]

12 minutes ago, mr moose said:

And what precise details where they that I supposedly claimed?

 

 

That they knew long before the general public that Zen was a 50+% increase over FX and that the specifics of Threadripper beyond being AMD's HEDT platform were known to them. None of which was apparent in their actions at the time.

[mr moose]

4 minutes ago, ravenshrike said:

That they knew long before the general public that Zen was a 50+% increase over FX and that the specifics of Threadripper beyond being AMD's HEDT platform were known to them. None of which was apparent in their actions at the time.

what I said was:

12 hours ago, mr moose said:

 

 

They knew.  They knew a lot more than you or I did and they know lot more than you or I think they know.

Which is a general statement which means regardless what you think you know, Intel know more.  It is not meant not be a specific claim of precisely what information they have and a concise list of accuracies.

 

 

[Stefan Payne]

34 minutes ago, mr moose said:

So let me get this straight, If I am accusing Intel of espionage you agree but if I am accusing AMD of it then I am making false statements? 

No, I don't agree on either side and say its bullshit that you accused either side of Espionage.

 

 

What you miss is that it was far simpler for AMD to guess what Intel might (or might not) do than the other way around.

Their "Tick-Tock" Strategy came back to bite them in the ass. 

 

Simple as that.

[mr moose]

3 minutes ago, Stefan Payne said:

No, I don't agree on either side and say its bullshit that you accused either side of Espionage.

 

 

What you miss is that it was far simpler for AMD to guess what Intel might (or might not) do than the other way around.

Their "Tick-Tock" Strategy came back to bite them in the ass. 

 

Simple as that.

Believe whatever you want.  It happens:

 

https://spectrum.ieee.org/semiconductors/processors/lessons-from-the-1-billion-intel-tradesecret-theft

[ravenshrike]

34 minutes ago, mr moose said:

what I said was:

Which is a general statement which means regardless what you think you know, Intel know more.  It is not meant not be a specific claim of precisely what information they have and a concise list of accuracies.

 

 

You were responding to a quote that stated they did not know that Zen was a 50% plus increase. You said they knew. That is a specific claim as to the extent of their industrial espionage capabilities.

[mr moose]

1 minute ago, ravenshrike said:

You were responding to a quote that stated they did not know that Zen was a 50% plus increase. You said they knew. That is a specific claim as to the extent of their industrial espionage capabilities.

Well, now we have that cleared up we can get on with the conversation.

[Stefan Payne]

20 minutes ago, mr moose said:

Believe whatever you want.  It happens:

https://spectrum.ieee.org/semiconductors/processors/lessons-from-the-1-billion-intel-tradesecret-theft

1 Nov 2008 | 4:00 GMT

Why you link to an old article from over 10 years ago? Why not link to newer sources that had the conclusion inside?! 

 

But why not show the whole story and link for example to Wikipedia, where there is more information:

https://en.wikipedia.org/wiki/Biswamohan_Pani

Quote

The information he apparently obtained is believed to be related to Intel's then next-generation Itanium microprocessor.

Öhm, yeah...

The Value for AMD I don't really see here...

What should AMD do with a failed Processor that never gained much relevancy??

At that time, Itanic should have been pretty dead...

Also:

Quote

On 8 August 2012 he was sentenced to three years in federal prison and given a fine of US$17,500.^[5]

Ähh...

That doesn't look that bad...

 

But hey, why not link to the Official Statement of the FBI:

https://archives.fbi.gov/archives/boston/press-releases/2012/former-intel-employee-sentenced-to-prison-for-stealing-valuable-computer-chip-manufacturing-and-design-documents

 

Because they stated these lines:

Quote

AMD also cooperated with the investigation, and there was no evidence that AMD or its employees had asked Pani to take these documents or even knew that he had them.

Yeah, right...

 

So you're saying that an Ex-Employee who didn't disclose that he was working for the Competition while Working for Intel, took some Documents to advance his career, without any evidence that AMD asked him to do it or even knew about it, is proof for what you claim it happens??
Yeah, right...

 

You know what:
It was the action of a not very nice person, who had no ethics and wanted to advance at all cost...

[Stefan Payne]

44 minutes ago, ravenshrike said:

The question is not does industrial espionage happen. It is whether Intel's sources were/are good enough to have the precise details that Moose is claiming.

Obviously not.

AMD has closed off every leak as did everyone.

And also its illegal with pretty draconian penalties from all sides.

 

Especially since its only possible in late stages, before that it might only be a couple of dozen or so people working on it.

44 minutes ago, ravenshrike said:

Given their reactions it's pretty clear that at best they only knew a month ahead of time for Zen and Threadripper as to the details of any announcements, and probably closer to a week beforehand.

Exactly, wich is not really espionage like Moose claims.

Especially since at that point a couple of thousand people know about the products, its not that secret anymore. So that is the point where Intel actually might have gotten some Infos -> Product validation phase from Partners.

 

If they did know about it sooner, the reaction Intel has shown, didn't show it. The ~1 Year for 6 Core Kaby-Lake also supports our theory than his because one year for a wider CPU is to be expected.

Especially since Intel didn't change much on the Cores itself. 

 

44 minutes ago, ravenshrike said:

It's also pretty clear that the leaks on the current architectures are all coming from manufacturing partners or people much further down the chain at AMD and that AMD is at least trying to be tightlipped about specifics even internally. The reason you didn't see a similar pattern for the initial Zen launch was because of how tight the launch window was. Remember all the initial difficulties in supply that resulted from the tight turnaround time motherboard manufacturers had.

Exactly.

Doesn't look too different with Matisse as well.

Rome is handed around a bit longer, so there Intel might know a bit more or so.

But even here, AMD might have had their personell around the chip and only showed it to a small amount of people who have deciding power and need to know.

[mr moose]

1 minute ago, Stefan Payne said:

1 Nov 2008 | 4:00 GMT

Why you link to an old article from over 10 years ago? Why not link to newer sources that had the conclusion inside?! 

 

But why not show the whole story and link for example to Wikipedia, where there is more information:

https://en.wikipedia.org/wiki/Biswamohan_Pani

Öhm, yeah...

The Value for AMD I don't really see here...

What should AMD do with a failed Processor that never gained much relevancy??

At that time, Itanic should have been pretty dead...

Also:

Ähh...

That doesn't look that bad...

 

But hey, why not link to the Official Statement of the FBI:

https://archives.fbi.gov/archives/boston/press-releases/2012/former-intel-employee-sentenced-to-prison-for-stealing-valuable-computer-chip-manufacturing-and-design-documents

 

Because they stated these lines:

Yeah, right...

 

So you're saying that an Ex-Employee who didn't disclose that he was working for the Competition while Working for Intel, took some Documents to advance his career, without any evidence that AMD asked him to do it or even knew about it, is proof for what you claim it happens??
Yeah, right...

So your arguing because he was only given 3 years that he didn't actually steal information from Intel?  or that this type of thing doesn't happen?

 

Are you just arguing for the sake of it? I mean it's pretty obvious that corporate espionage happens (I have linked to proof), it is also not up for debate that all tech companies keep a close eye on their competition, sometimes through leaks and secret deals, sometimes through old fashioned sleuthing.   I'm not sure why this is a hard thing to accept?

[JoostinOnline]

2 hours ago, Master Disaster said:

Well you're the only person here naive enough to not believe it happens. 

Where did I say that?

[r2724r16]

1 hour ago, Schnoz said:

once you bring a 135-watt 16-core 5.1 GHz processor to the market on a motherboard that won't have to be upgraded, Intel is screwed.

Not happening anytime soon. I don't care what anyone tells me.

[Stefan Payne]

2 minutes ago, mr moose said:

So your arguing because he was only given 3 years that he didn't actually steal information from Intel?  or that this type of thing doesn't happen?

No, it was a sick individual who stole documents for his own benefit but without benefitting anyone or the knowledge of another company.

 

All that it shows is that the Security measures at Intel at the time were shit.

And that that person was a garbage individual.

[mr moose]

1 minute ago, Stefan Payne said:

No, it was a sick individual who stole documents for his own benefit but without benefitting anyone or the knowledge of another company.

 

All that it shows is that the Security measures at Intel at the time were shit.

And that that person was a garbage individual.

Head in the sand or fingers in the ears, whatever...

 

[GoldenLag]

7 minutes ago, r2724r16 said:

Not happening anytime soon. I don't care what anyone tells me.

I mean, we have a 8 core 180-230(140-180) watt 5-5,1ghz CPU at 14nm.

 

Just because it says 135 watts doesnt mean its gonna pull that ammount of power, nor does it mean every core will run at that speed. 

[r2724r16]

Just now, GoldenLag said:

I mean, we have a 8 core 180-230(140-180) watt 5-5,1ghz CPU at 14nm.

  

Just because it says 135 watts doesnt mean its gonna pull that ammount of power, nor does it mean every core will run at that speed. 

What's your point?! I'm skeptical about the part that says that Zen 2 chips will have up to 16 cores @ 5.1 GHz.

[Stefan Payne]

1 minute ago, mr moose said:

Head in the sand or fingers in the ears, whatever...

No, because he could for example have used the Documents for Blackmail, to get a bit more money out of Intel.

In that case it wouldn't be Espionage, would it?

 

What he did was stealing documents and violated company policies.

But with those Documents, it seems there wasn't anything done.

He just had them. 

AMD worked with the FBI, they couldn't find any evidence that AMD knew about the Documents...

 

[GoldenLag]

2 minutes ago, r2724r16 said:

What's your point?! I'm skeptical about the part that says that Zen 2 chips will have up to 16 cores @ 5.1 GHz.

Untill launch day we cant be sure what we will get. We will have to see. Though 5,1ghz effective speed on 16 cores would not leVe the realm of possibility

[Stefan Payne]

3 minutes ago, r2724r16 said:

What's your point?! I'm skeptical about the part that says that Zen 2 chips will have up to 16 cores @ 5.1 GHz.

https://www.anandtech.com/show/13562/amd-announces-radeon-instinct-mi60-mi50-accelerators-powered-by-7nm-vega

Scroll down, the "World's first 7nm GPUs" Part.

>1.25x higher performance at the same power.

Here something similar:

https://liliputing.com/2018/06/amds-first-7nm-gpu-coming-this-year-7nm-cpu-in-2019.html

 

max. Boost Clock of Ryzen 2700x is 4,3GHz. 

4,3*1,25 -> 5,375GHz

 

Even +15% higher clockrates than max Boost of Ryzen 2700X means 5GHz.

[r2724r16]

Just now, Stefan Payne said:

https://www.anandtech.com/show/13562/amd-announces-radeon-instinct-mi60-mi50-accelerators-powered-by-7nm-vega

Scroll down, the "World's first 7nm GPUs" Part.

>1.25x higher performance at the same power.

Here something similar:

https://liliputing.com/2018/06/amds-first-7nm-gpu-coming-this-year-7nm-cpu-in-2019.html

 

max. Boost Clock of Ryzen 2700x is 4,3GHz. 

4,3*1,25 -> 5,375GHz

 

Even +15% higher clockrates than max Boost of Ryzen 2700X means 5GHz.

As much as I'd like to believe all of this, I just don't think it's gonna happen. Especially at the rumoured pricepoints (~U$500 for the 16 core CPU).